Skip to main content
. 2021 Aug 18;4(8):e2121410. doi: 10.1001/jamanetworkopen.2021.21410

Table 2. Out-of-Pocket Spending by Coinsurance Percentage and Timing of Search, 2015-2016a.

Spending Percentage of coinsurance and timing of search
1%-5% (n = 15 687) 6%-10% (n = 30 361) ≥11% (n = 47 458)
Never Early Late Never Early Late Never Early Late
Mean out-of-pocket spending, $ 623.66 681.96 766.46 1446.45 1529.12 1573.09 2486.97 2560.36 2592.16
% Change in out-of-pocket spending (95% CI) 0 [Reference] 6.8 (1.1-12.8) 14.5 (6.6-22.9) 0 [Reference] 4.3 (2.4-6.1) 5.8 (3.4-8.3) 0 [Reference] 2.4 (1.3-3.5) 2.9 (1.3-4.6)
Mean difference (95% CI), $ 0 [Reference] 42.35 (6.64- 80.08) 90.38 (41.43-142.94) 0 [Reference] 61.74 (35.39-88.57) 83.75 (48.93-119.38) 0 [Reference] 59.57 (33.44- 85.96) 73.33 (32.04-115.29)
P valueb <.001 <.001 <.001 <.001 <.001 <.001
a

Model results are from linear regression models as described in the Methods section. Delivery episodes with missing values for area median income were excluded from the analysis. Likelihood ratio tests failed to reject the null hypothesis that the coefficients on early and late searchers were equal. The percentage change in out-of-pocket spending was calculated by exponentiating the coefficient from the model and converting it into a percentage; the dollar change multiplied that percentage by the mean out-of-pocket spending in the never searcher group.

b

P values were relative to the reference category (never searcher).